Ashwani Damodaran, Lead Architect for Data Repositories Management, Societe Generale’s Global Solution Centre writes about datawarehouse migration and the factors to consider before migrating your applications.
The evolution of Data
Data has evolved over the last five decades owing to the transformation of technology itself. Everyday data is generated at a rate of 1.5 megabytes per second and it does not require a computer! Here are some common activities that result in this large amount of data being produced:
To manage such large volumes of data being produced, database management systems came into existence such as:
- Database Management System (DBMS)
- Relational Database Management System (RDBMS)
- Online Transactional Processing (OLTP)
- In memory databases
- Data Warehouse
- NoSQL databases
Emerging technologies from data evolution
Experts began looking at whether there is value the information being generated. In general about 80% of the data produced is not used. The bigger challenge was extracting value at a rate that was feasible to businesses. Thus big data was born. Furthermore, the ability to retain volumes of data also supports Artificial Intelligence and Machine Learning leading to multiple solutions based on emerging technology.
Is big data meant for your application?
Big data has taken over the industry. But is it worth implementing it just for the sake of it? Here are some details to research before you consider big data as a solution:
- How much volume are you dealing with? – Big data is not recommended for the application dealing with low volume of data.
- What is the velocity of your data? – If your application is aimed towards dealing with high velocity of data and analysis time is critical, then your application is a candidate for big data.
- Do you have variety of data? – Big data solutions are considered as a good fit for applications which are meant to deal with different types of data.
- How much veracity does your data carry? – Uncertainty in data and data quality may be due to manual action, bugs, data transfer issues (noise) and so on. In such cases it is better to consider big data solutions.
If you feel that these points suit the application, then you can consider big data to also modernise applications. This brings us to the next parameter to consider:
Should existing applications migrate to big data platform?
Migration is challenging and you should consider these aspects before application modernisation:
- There should be definitive migration strategy considering cost and benefit analysis
- Evaluate if the migration tools can help you to upgrade
- Testing strategy should be built to ensure successful migration
- Rely on unit testing and regression testing tools
- New system should be scalable, reliable, maintainable and should have considerable performance improvement
- Plan to migrate the current system to a new system ensuring no data loss/corruption
Last but not the least, infrastructure – on premises or public cloud?
- Cloud PAAS is more cost effective than on premises
- Data confidentiality is a concern with public cloud. Adaptation strategy should include security measures
- Ensure that on premises applications can seamlessly interact with public cloud – refactor code and manage a demand to rearchitect your solution
The decision of application modernisation should be driven by need. While the above parameters might seem easy to consider, it’s recommended to allow experts to manage the infrastructure ensuring that the goal of the application is not affected.
hours minutes seconds
Virtual Meetup: Future of Data Transformation